Indexing iconic image database for interactive spatial similarity retrieval

Similarity-based retrieval of images is an important task in many image database applications. Interactive similarity retrieval is one way to resolve the fuzzy area involving psychological and physiological factors of individuals during the retrieval process. A good interactive similarity system is not only dependent on a good measure system, but also closely related to the structure of the image database and the retrieval process based on the respective image database structure. In this paper, we propose to use a digraph of most similar image as an index structure of an iconic spatial similarity retrieval. Our approach makes use of them simple feedback from the user, and avoids the high cost of recomputation of interactive retrieval algorithm. The interactive similarity retrieval process is similar to a guided navigation by the system measure and the user in the image database. The proposed approach prevents looping and guarantees to find the target image. It is straightforward and adaptive to different similarity measure.